Move from another app

Turn the export file I've attached into a CSV that the Resurf app can import. If you can create files, give me a file called captures.csv. Otherwise reply with only the CSV in one code block.

Follow these rules exactly:

1. The first line is exactly:
type,title,source_url,description,user_note,tags,spaces,created_at,pinned_at
2. Write one row for every saved item that has a web address starting with http:// or https://. Skip folders, separators, bookmarklets and items without a web address, and tell me after the CSV how many you skipped.
3. type: always link, in lower case.
4. title: the item's title or name. Empty if it has none.
5. source_url: the item's full web address, exactly as in the file.
6. description: the excerpt or summary the app saved for the item. Empty if none.
7. user_note: my own note or comment on the item, including any highlights. Empty if none.
8. tags: the item's tags plus the name of the folder, collection or board it was in, separated by semicolons (;) with no spaces around them, for example design;inspiration;Travel. Don't add generic folders like "Bookmarks bar", "Unsorted" or "Other bookmarks". Empty if none.
9. spaces: leave empty.
10. created_at: when the item was saved, in UTC, as YYYY-MM-DDTHH:MM:SSZ, for example 2024-03-01T09:30:00Z. Convert Unix timestamps (seconds, milliseconds or microseconds) and written dates. If only the day is known, use T00:00:00Z. If no date is known, leave it empty.
11. pinned_at: if the item was a favorite, starred or pinned, copy its created_at here. Otherwise empty.
12. Separate columns with commas. Wrap any value that contains a comma, a double quote or a line break in double quotes, and write each double quote inside it twice.
13. Keep every item and keep the original words. Don't invent, translate or shorten anything.

If there are more than 500 items, split the CSV into parts of 500 rows, each starting with the same first line, and give me one part at a time.

import csv, sys
from datetime import datetime, timezone

def iso(value):
    if not value:
        return ""
    when = datetime.fromisoformat(value.replace("Z", "+00:00"))
    return when.astimezone(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")

columns = ["type", "title", "source_url", "description", "user_note", "tags", "spaces", "created_at", "pinned_at"]
with open(sys.argv[1], newline="", encoding="utf-8-sig") as source, \
     open("captures.csv", "w", newline="", encoding="utf-8") as target:
    writer = csv.DictWriter(target, fieldnames=columns)
    writer.writeheader()
    for row in csv.DictReader(source):
        created = iso(row.get("created", ""))
        tags = [tag.strip() for tag in row.get("tags", "").split(",") if tag.strip()]
        writer.writerow({
            "type": "link",
            "title": row.get("title", ""),
            "source_url": row.get("url", ""),
            "description": row.get("excerpt", ""),
            "user_note": row.get("note", ""),
            "tags": ";".join(tags),
            "spaces": row.get("folder", ""),
            "created_at": created,
            "pinned_at": created if row.get("favorite") == "true" else "",
        })

import csv, sys
from datetime import datetime, timezone
from html.parser import HTMLParser

class Bookmarks(HTMLParser):
    def __init__(self):
        super().__init__()
        self.folders, self.pending, self.rows, self.link, self.text = [], None, [], None, ""

    def handle_starttag(self, tag, attrs):
        attrs = dict(attrs)
        if tag == "dl":
            self.folders.append(self.pending or "")
            self.pending = None
        elif tag in ("a", "h3"):
            self.link = attrs if tag == "a" else None
            self.text = ""

    def handle_data(self, data):
        self.text += data

    def handle_endtag(self, tag):
        if tag == "h3":
            self.pending = self.text.strip()
        elif tag == "dl" and self.folders:
            self.folders.pop()
        elif tag == "a" and self.link and self.link.get("href", "").startswith("http"):
            added = self.link.get("add_date", "")
            created = datetime.fromtimestamp(int(added[:10]), timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ") if added.isdigit() else ""
            folders = [name for name in self.folders if name]
            self.rows.append({"type": "link", "title": self.text.strip(), "source_url": self.link["href"],
                              "tags": ";".join(folders), "created_at": created})
            self.link = None

parser = Bookmarks()
with open(sys.argv[1], encoding="utf-8") as source:
    parser.feed(source.read())
with open("captures.csv", "w", newline="", encoding="utf-8") as target:
    writer = csv.DictWriter(target, fieldnames=["type", "title", "source_url", "tags", "created_at"])
    writer.writeheader()
    writer.writerows(parser.rows)
print(f"{len(parser.rows)} bookmarks written to captures.csv")

import csv, json, sys

with open(sys.argv[1], newline="", encoding="utf-8-sig") as source:
    rows = [row for row in csv.DictReader(source) if row.get("Highlight", "").strip()]

items = []
for number, row in enumerate(rows, start=1):
    author, book = row.get("Book Author", "").strip(), row.get("Book Title", "").strip()
    item = {
        "id": str(number),
        "intent": "quote",
        "content": {"kind": "text", "text": row["Highlight"].strip()},
        "metadata": {"quoteAttribution": ", ".join(part for part in (author, book) if part)},
        "organization": {"tags": [tag.strip() for tag in row.get("Tags", "").split(",") if tag.strip()]},
    }
    if row.get("Note", "").strip():
        item["organization"]["userNote"] = row["Note"].strip()
    items.append(item)

for start in range(0, len(items), 100):
    page = start // 100 + 1
    with open(f"readwise-{page}.json", "w", encoding="utf-8") as target:
        json.dump({"requestID": f"readwise-import-{page}", "items": items[start:start + 100]}, target, ensure_ascii=False)
    print(f"readwise-{page}.json")